The Journal of Neuroscience, January 1, 1996, 16(1):130-136

Rise in Intracellular Calcium via a Nongenomic Effect of Allopregnanolone in Fetal Rat Hypothalamic Neurons

Govindan Dayanithi and Lucia Tapia-Arancibia

Laboratoire de Neurobioloaie Endocrinolouiaue, URA I 197 CNRS, Universit6 Montpelier II, F-34095 Montpelier, Cedex 5, France

This study examines the early effects of 3a-hydroxy-%-preg- nan-20-one (allopregnanolone) on cytosolic free calcium con- centration ([Ca2’],) in primary cultures of fetal rat hypothalamic neurons. Microspectrofluorimetry of fluorescent Ca2+-sensitive indicator Fura- was used to quantify these changes. Allopreg- nanolone (1 PM to 100 nM) increased [Ca2’], within 2-3 set, in a dose-dependent manner, with an EC,, of 10 5 4 nM. The stimulatory effect of allopregnanolone was attributable princi- pally to a Ca2+ influx, as shown by the strong inhibition of external Ca”+ removal or by the calcium channel blocker nifed- ipine. The effect was stereospecific because the allopreg- nanolone isomer 3P-hydroxy-5a-pregnan-20-one had no effect

on [Ca”],. Among two other steroids examined, progesterone had no effect on [Ca2’li, but 17P-estradiol evoked a rise in [Ca2’],, although to a lesser extent than allopregnanolone. The allopregnanolone-induced [Ca2’], rise was inhibited by picro- toxin and bicuculline but was unaffected by tetrodotoxin or by pretreatment of neurons with pertussis toxin. These results are consistent with a membrane site of action for allopregnanolone associated with GABA, receptors, leading to rapid changes in [Ca”], in fetal rat hypothalamic neurons.

Key words: neurosteroids; calcium; hypothalamic neurons; Fura-2; allopregnanolone; nongenomic effect; GABA, receptor

The term “neurosteroids” was introduced by Baulieu (1981) to describe the steroids synthesized in brain tissue from cholesterol or from steroid hormone precursors (Robe1 and Baulieu, 1990, 1994; Baulieu, 1991; Orchinick and McEwen, 1993). According to their effects, they can be considered “neuroactive steroids.” This term, coined by Paul and Purdy (1992), includes all steroids that are active on neuronal tissue regardless of whether they are synthesized in the brain.

Aside from the classical genomic actions of steroids (McEwen, 1991a), these substances may alter rapidly the excitability of neurons. In fact, the ability of certain steroids and their metabo- lites to influence brain activity, including firing rate of neurons, induction of sedation, anesthesia, neurosecretion, and behavioral changes, is well known (Schumacher, 1990; McEwen, 1991b). However, the mechanisms involved in these actions are not well understood. Recently, Ramirez et al. (1990) demonstrated that progesterone and its metabolites can modify neuronal secretion by acting through a nongenomic transmembrane-signaling mech- anism. They postulated that progesterone could open Ca*+ chan- nels of nerve terminals leading to a rapid neuronal secretion. Evidence suggesting a role of CazC in this phenomenon was provided previously by Drouva et al. (1985), who showed that omission of Ca*+ in the medium or addition of D-600, a Ca2+ channel blocker, antagonized the stimulatory effect of progester-

Received June 19, 1995; revised Sept. 15, 1995; accepted Sept. 26, 1995. This work was supported by grants from CNRS and INSERM (CRE 910701). We

thank Prof. Jorae Belmar, Prof. HBl&ne Astier, and Dr. Cohn Ingram for valuable comments and critical reading of the manuscript, and Mr. Edm&d &vary for his technical assistance. We also thank Roussel UCLAF, Romainville, France for the eenerous eift of RI1 16709.

Correspondence should be addressed to Dr. L. Tapia-Arancibia, Laboratoire de Neurobiologie Endocrinologique, URA 1197 CNRS, Universite Montpellier II, Place Eugtne Bataillon, F-34695 Montpellier, Cedex 5, France. Copyright 0 1995 Society for Neuroscience 0270-6474/95/160130-07$05.00/O

one on luteinizing hormone-releasing hormone release from me- diobasal hypothalamic slices.

A mechanism involved in the nongenomic effect of steroids involves binding to a membrane-bound receptor complex, such as those for inhibitory, e.g., GABA (Majewska et al., 1986), and/or excitatory, e.g., NMDA neurotransmitters (Smith, 1987; Paul and Purdy, 1992; Mellon, 1994). Some steroids may potentiate or inhibit NMDA receptor-mediated responses (Smith et al., 1987; Wu et al., 1991; Irwin et al., 1992, 1994; Park-Chung et al., 1994), and they can also have agonistic or antagonist GABA, receptor- like activity (Paul and Purdy, 1992).

Interestingly, two of the neuroactive steroids that are the most potent negative modulators of GABA, receptors, 3a- hydroxy-5a-pregnan-20-one (allopregnanolone), which is the primary metabolite of progesterone, and allotetrahydrodeoxy- corticosterone, derived from deoxycorticosterone, have been measured in brain and plasma, where their levels fluctuated in response to stress and during the estrus cycles of female rats (Paul and Purdy, 1992). In normal male rats, for example, allopregnanolone is detectable in the cerebral cortex and hy- pothalamus, its level being markedly and rapidly increased after acute swim stress (Purdy et al., 1991) or acute electro- shock (Korneyev et al., 1993).

Taken together, these data strongly suggest that neuro- steroids represent a new class of neuromodulators that can rapidly alter neuron excitability via nongenomic mechanisms. However, little information is available on the molecular mech- anisms, particularly the excitatory ones, that may explain the rapid membrane effects of steroid hormones. Consequently, the aim of the present study was to examine the effect of allopregnanolone on intracellular calcium level ([Ca’+],) in primary cultures of rat hypothalamic neurons by performing high-time resolution photometric measurements of [Ca2+li transients evoked by drug application.

Dayanithi and Tapia-Arancibia l Allopregnanolone and [Ca’+], Rise in Hypothalamic Neurons J. Neurosci., January 1, 1996, 76(1):130-136 131

MATERIALS AND METHODS Chemicals. Most of the standard chemicals, tetrodotoxin (TTX), EGTA, 17P-estradiol, progesterone, pertussis toxin (PTX), and nifedipine, were obtained from Siema ISt. Louis. MO\. Fura- AM and Pluronic F-127 were purchased -from Molecular Piobes (Eugene, OR). Allopreg- nanolone (3a-hydroxy-5a-pregnan-20-one or 3ol,Sa-THP or RU 16709) was kindly provided by Roussel UCLAF (Romainville, France)

Cell culture. Primary cultures were prepared by mechanoenzymatic dissociation of fetal (day 17) Sprague-Dawley rat hypothalami as de- scribed previously (Tapia-Arancibia et al., 1988) with necessary modifi- cations. Briefly, cells were plated onto chambered cover glass (Lab-Tek N” 178565, Nunc, Naperville, IL) coated previously with poly-D-lysine (10 &ml; M, 220 kDa) and preincubated for 1 hr with 10% fetal calf serum in Minimum Essential Medium (MEM; Gibco, Grand Island, NY). Cells were seeded in growth medium composed of MEM supplemented with Nu serum (10%) (Collaborative Research, Lexington, MA), glucose (0.6%), glutamine (2 tnM), and penicillin-streptomycin (2.5 U/ml), ad- justed to pH 7.3. The cultures were maintained at 37°C in a humid atmosphere (95% sir/5% CO,). The proliferation of non-neuronal cells was inhibited by treatment with 10 pM cytosine-arabinoside for 48 hr between days 3 and 5 after plating. Three days before the experiments, a phenol red-free medium supplemented with 10% charcoal-treated Nu serum (devoid of steroids) was used. Experiments were performed after 8 d in culture.

Measurements of intracellularfree calcium in single neurons. The [Ca”], was determined from the intensity of the fluorescence signal with the Cazc-sensitive indicator Fura-2. Hypothalamic neurons in culture were washed with a serum-free medium (Locke’s buffer) containing (in ITIM):

NaCl 140, KC1 5, glucose 10, KH,PO, 1.2, MgSO, 1.2, CaCl, 1.8, HEPES 10, adjusted to pH 7.4 with NaOH. The osmolarity of the buffer was 300-310 mOsm/kg. Cells were loaded with Fura- AM, which was dis- solved first in water-free dimethyl sulfoxide (DMSO) and then diluted with Locke’s buffer to a final concentration of 2.5 pM Fura- AM, which containing 0.02% (w/v) Pluronic F-127. The DMSO concentration in these mixtures was <O.Ol%. Dye loading was carried out for 30 min in a humidified O&O, incubator at 37°C. Subsequently, loaded cells were washed with Locke’s buffer and the cover glass was mounted on the microscope stage. The fluorescence measurements were performed with buffer and drug solutions that were maintained at 36-37°C throughout the measuring period.

Fast fluorescence photometer system. [Ca2+li in single cells was mea- sured by the method described by Lambert et al. (1994) with necessary modifications. Briefly, the Zeiss (Oberkochen, Germany) photometer system was based on an inverted microscope (Axiovert 100) equipped for epifluorescence. Interference filters of 340/10 and 380/10 nm were mounted alternately (200 Hz) on the filter wheel, and excitation light was deflected through an oil-immersion objective (Zeiss-plan Neoflaur 100X 1.30, Ph 3). Fluorescence emission from individual cells was viewed through a diaphragm adjusted to the cell size. To minimize the back- ground noise of the Fura- signals, successive values were averaged to a final resolution of 320 msec. With fluorescence values corrected for background and dark current, calculation of [Ca”], was carried out from the ratio between 340 and 380 nm recordings, in accordance with the equation given by Grynkiewicz et al. (1985): [Ca2+li = Kd X (R -

Rnd(%ax - R) X /3. To calculate [Ca2+li values from measured data, the parameters Kd = 224, Rmin = 0.203, R,,, = 6.351, and p = 3.418 were specified in the calibration menu.

Solutions and drug application. Stock solutions (lo-* M) of the steroids dissolved in ethanol were prepared daily and then diluted with Locke’s buffer containing TTX (1 PM). All steroid solutions (10 ~1) were applied within 5 set directly in the proximity of the recorded neurons (10 yl in a total volume of 200 ~1) using a motor-driven syringe (Hamilton-microlab 900, speed setting 9; Bonaduz, Switzerland). The final concentration of ethanol in these solutions was <O.Ol%. Ethanol (0.01%) has no effect by itself on basal [Ca’t]i. This method allowed reproducible changes in the medium surrounding a selected neuron and appeared to ensure fast concentration changes. Calcium-free medium (EGTA-buffer) contained (in mM): EGTA 2, NaCl 140, KC1 5, glucose 10, KH,PO, 1.2, MgSO, 1.2, and HEPES 10, pH 7.4. In this EGTA buffer, 100 nM free Ca2* (which corresponds to the resting [Ca2+li as determined with Fura- measure- ments) was maintained.

Data analysis. Results were expressed as mean 2 SEM. The statistical evaluation was performed by Student’s t test. The estimations of EC,,, and its standard error were determined with a Marquardt-Levenberg nonlin-

ear least-squares curve-fitting algorithm (Origin, Microcal Software, Northampton, MA).

RESULTS

Effect of allopregnanolone on intracellular calcium The basal level of [Ca2+li in hypothalamic neurons was 71 -+ 8 nM (mean t SEM, n = 21).

Applications (60-120 set) of allopregnanolone (1 PM to 100 nM) evoked a marked increase of [Ca”]i, reaching a peak rise within 2-3 set in 85 of 113 neurons tested. Two distinct types of [Ca’+], responses were observed: (1) in a great majority of neu- rons (67/85, 79%), allopregnanolone induced a brief, single, tran- sient [Caztli rise that decayed to near resting levels in ~30 set in the continued presence of the steroid (see, for example, 10 and 100 nM in Figs. L4, 2, 3B, 4, and 6B); and (2) allopregnanolone triggered a single, sustained rise in [Ca2+li (18/85 neurons, 21%) with a plateau phase, decaying very slowly throughout the mea- suring period (see, for example, 1 nM in Figs. lA, 3A, and 6A).

Figure L4 shows the increases in [Ca2+li induced by increasing concentrations of allopregnanolone. In six neurons tested, a sig- nificant rise in [Cazcli is observed by steroid concentrations as low as 1 PM. Allopregnanolone elicits a concentration-dependent in- crease in [Ca’+], within the range of 1 PM to 100 nM. Maximal response is obtained at 100 nM, and the estimated EC,,, value for allopregnanolone is 10 ? 4 nM (Fig. 1B).

Successive applications of allopregnanolone evoked a similar [Ca2+li rise. Figure 2 shows the [Caz+li rise in response to two repetitive applications of 1 nM allopregnanolone. Four neurons tested displayed similar profiles. This result allowed us to use for some experiments the same neuron as its own control, thus diminishing the effect of individual variation among neurons.

Blockade of allopregnanolone-induced changes in [Ca’+]i To investigate whether the allopregnanolone-induced [Ca2’li rise was dependent on the presence of external Ca*+, allopreg- nanolone-sensitive neurons that showed a rise in [Ca2+li in the presence of 1.8 mM Ca*+ were treated with the steroid in EGTA buffer. Figure 3A shows that the allopregnanolone-induced [Ca2+li rise was reduced significantly in the presence of EGTA buffer. In the presence or absence of Ca*+ in the external me- dium, the [Ca2+li rise induced by 10 nM allopregnanolone was 466 2 32 and 178 ? 10 nM, respectively (p < 0.01, n = 9). As shown in Figure 3A (right), there was still a small residual response to allopregnanolone in the absence of external Ca’+, and this was observed consistently in all nine neurons tested. To examine whether the Ca2’ influx could involve voltage-sensitive Ca2+ channels, we tested the L-type, dihydropyridine-sensitive Ca2+ channel blocker nifedipine (Fig. 3B). Nifedipine (10 PM) applied 30 set before and during the application of 10 nM allopreg- nanolone diminished by 75 -C 12% (n = 8) the calcium rise induced by allopregnanolone: 133 2 23 vs 545 t- 43 nM (control) (p < 0.01, n = 8). These experiments, with or without nifedipine, were performed on the same neurons.

Specificity of allopregnanolone effect on [Ca2+]i Dvo different approaches were used to define pharmacologically the steroid specificity. First, we investigated the ability of two other steroids (progesterone and 17/3-estradiol) to increase [Ca2+li in the hypothalamic neurons. Progesterone (10 nM to 1 FM) had no signif- icant effect on [Ca2+li. In these experiments, the mean [Ca2’li values obtained in the absence or presence of 10 nM progesterone were 89

132 J. Neurosci., January 1, 1996, 76(1):130-136 Dayanithi and Tapia-Arancibia . Allopregnanolone and [Ca*+], Rise in Hypothalamic Neurons

F&re 1. Dose-dependent effect of allopregnanolone (3+5a-THP). A, Truces represent changes in [Ca2’], induced by increasing concentrations of allopreg- nanolone (0.001-100 nM) measured in a single neuron. Each stimulus was separated by 5 min intermission after washing twice in normal Locke’s buffer. B shows the dose-response relationship of evoked [Ca”], in- crease (mean t SEM) as a function of allopreg- nanolone concentration. Data were obtained from sev- eral neurons (sample size is given in brackets) for a given concentration tested randomly. *p < 0.01; **p < 0.001 vs control without allopregnanolone.

A 750 -

0.001 nM 0.01 nM 0.1 nM 500 -

OJ , 60 (a)

B

0 Basal -12 -11 -10 -9 -8 -7 -6 -5

3a, 5a-THP (log M)

+ 9 and 106 ? 5 nM, respectively (p > 0.05, rz = 7). Similarly, progesterone applied at higher concentration did not significantly affect the [Ca”‘], levels (p > 0.05, n = 6), which ranged between 79

180,

3a,5a-THP (1 nM)

3a, 5a-THP (1 nM) I

-

0 60

Time (s)

120

Figure 2. Successive allopregnanolone (3a,5a-THP) applications on [Ca’+], increase. A selected neuron was subjected to two successive applications of allopregnanolone (1 nM). The resting [Ca”‘], had returned to prestimulation level before the second application. Note that the amplitude of [Ca2’], response induced by allopregnanolone is unchanged. Four neurons displayed similar profiles.

t 16 nM (control) and 88 -C 13 nM (1 pM progesterone). Interest- ingly, these nonresponding neurons to progesterone exhibited a remarkable [Ca’+], rise to 10 nM allopregnanolone (mean peak value: 533 + 96 nM, rz = 6). In contrast, 17/%estradiol evoked [Ca2’li increase to a lesser extent than allopregnanolone: 17@estradiol(lOO nh?)-induced [Ca”‘], rise peaked at 288 -C 56 nM (n = 10) vs 1022 t 20 nM (n = 8) (p < 0.01) induced by 100 nM allopregnanolone. The second paradigm consisted of examining the effect of the P-hydroxy- lisomer of allopregnanolone. Figure 4 shows that 3/3-hydroxy-Sol- pregnan-20-one used at the same concentration of that of the 3~ isomer (1 nM) had no effect on [Ca2’],. However, after washing in Locke buffer the same neuron showed a marked response to allo- pregnanolone. The resting [Ca2+li was unchanged after addition of 3/3-hydroxy-5oc-pregnan-20-one (68 2 7 nM>, whereas in the same neurons 3a-hydroxy-5ol-pregnan-20-one augmented [Caz’li up to 327 + 37 nivr (n = 7).

Lack of effect of PTX on allopregnanolone-induced [Ca*‘], Hypothalamic neurons were incubated for 20 hr with 100 n&ml PTX before loading with Fura- AM. [Ca”‘], measurements also were carried out in the presence of PTX, which had no significant effect (p > 0.05) on allopregnanolone (10 nrvl)-induced [Ca”‘], rise (Fig. 5).

Dayanithi and Tapia-Arancibia l Allopregnanolone and [Ca”], Rise in Hypothalamic Neurons J. Neurosci., January 1, 1996, 16(1):130-136 133

A EGTA buffer

I I 3a, 5cx-THP (10 nM) 3u, 5wTHP (10 nM)

400 1 ,

I

-

0-l I 60s ’

B

600 3a, !?a-THP (10 nM) nifedipine (10 I.IM)

1 ’ II 1 I

3u, 5c(-THP (10 nM)

I , I

Figure 3. Effect of external CaZ’ on [Ca’+], increase induced by allopregnanolone (301,5c~-THP). A, The left panel represents the [Ca’+], response induced by 10 nM allopregnanolone in the presence of Locke’s buffer containing 1.8 mM external Ca*+. Preincubation of the same neuron in low Ca*+-EGTA buffer (see Materials and Methods) significantly abolished [Ca”], increase induced by allopregnanolone (right). Nine neurons displayed similar profiles when tested under similar conditions. B, Effect of L-type channel blocker, nifedipine, on [Ca’+], increase induced by allopregnanolone (3a,5wTZfP). Left, Con- trol [Ca’+], response induced by 10 nM allopregnanolone. Right, After washing, the same neuron was preincubated for 30 see with 10 pM nifedipine and then subjected to 10 HIM allopregnanolonc. The [Ca’*], response induced by allopregnanolone is markedly decreased by the L-channel blocker. Nine neurons displayed similar profiles when tested under similar conditions.

250

52

$ 125 “m 2

0

Inhibitory effect of picrotoxin and bicuculline on allopregnanolone-induced [Ca”], In an attempt to determine whether the [Ca”], rise induced by allopregnanolone involved GABA, receptor participation, we performed a series of experiments with and without picrotoxin or bicuculline (Fig. 6) using the same neurons as their own control. Picrotoxin (10 pM) applied simultaneously with 10 nM allopreg- nanolone inhibited by 65 t- 7% the intracellular Ca2+ response to allopregnanolone: 422 5 41 vs 1.50 ? 27 nM (p < 0.01, IZ = 11) (Fig. 6A). Bicuculline (50 FM) applied simultaneously with 10 nM allopregnanolone completely blocked (Fig. 6B) the peak of [Cazfli rise induced by allopregnanolone: 434 + 55 vs 47 2 5 nM (n = 8) (Fig. 6B).

DISCUSSION This is, to our knowledge, the first study showing very rapid stimulatory effect of a neurosteroid, allopregnanolone, on cytoso- lit free Ca2+ in central nervous system neurons. The peak was reached within 2-3 set after allopregnanolone application and is one of the most rapid membrane effects of a steroid to be reported. It may be comparable with that observed for the acti- vation of calcium channels by progesterone in plasma membranes of human sperm (Blackmore et al., 1990), or with the effect of steroids on electrical activity of nerve cells occurring when applied locally (Robe1 and Baulieu, 1990; McEwen, 1991a).

Our results suggest that allopregnanolone-induced [Ca’*], rise is dependent mainly on the presence of extracellular Ca’+. In- deed, EGTA buffer strongly reduced the increased response to allopregnanolone. In addition, -70% of the allopregnanolone- induced increase in [Ca”], is inhibited by nifedipine, suggesting that part of the calcium influx occurs via voltage-gated calcium channels of L-type, which have been described in these cells (Tapia-Arancibia and Humbert, 1991; Desarmenien et al., 1994; Dayanithi et al., 1995). These experiments also clearly show that there is still a residual [Ca”], response to allopregnanolone. This response may be attributable to the involvement of intracellular Ca2+ stores (Simpson et al., 1995) or to the involvement of other voltage-activated Ca2+ channel types (Hofman et al.! 1994). Fur- ther experiments are necessary to characterize this residual re- sponse to allopregnanolone. However, GABA, receptors seem to be involved in the [Cazili rise induced by allopregnanolone, as suggested by the inhibitory effect of picrotoxin and bicuculline.

3p, 5a-THP (1 nM) 3~r, 5~THP (1 nM) I I I I

--

I 90 s I

L Figure 4. Specificity of allopregnanolone (345a-THP) on [Ca’+], rise. Left, A selected neuron was subjected first to the isomer 3&5a-THP (1 IIM). Rig&, After washing, the same neuron exhibited a marked [Ca’+], increase when the isomer 3c1,5ol-THP (1 nM) was applied. Seven neurons displayed similar profiles.

134 J. Neurosci., January 1, 1996, 76(1):130-136 Dayanithi and Tapia-Arancibia l Allopregnanolone and [Ca”], Rise in Hypothalamic Neurons

0

Figure 5. Lack of effect of PTX on 3qSol-THP-induced [Ca*+], rise. The culture dishes were incubated for 20 hr with (n = 8) or without (n = 8) 100 ngiml PTX and then were tested with 10 nM allopregnanolone. Diagram represents basal levels of [Ca*‘], and allopregnanolone-evoked [Ca2’], increase in control neurons, and the response to allopregnanolone in the presence of PTX in PTX-treated neurons.

This is consistent with a potentiation by allopregnanolone of a GABA,-mediated depolarization of neurons that secondarily may allow opening of voltage-sensitive calcium channels. We have shown previously, using the same experimental model, that GABA is released tonically in the incubation medium from hy- pothalamic neurons (Rage et al., 1992, 1993). According to this interpretation, it has been shown that in developing cortical neu- rons (Yuste and Katz, 1991) in cultured cerebellar granule cells (Connor et al., 1987), and in hippocampal neurons in the early postnatal life (Cherubini et al., 1991) GABA is able to depolarize cells.

The allopregnanolone-induced [Ca2+], rise is dose-depen- dent, saturable, stereoselective, and specific because it is not mimicked by the 3P-hydroxy isomer or by progesterone. A

A

s s $44 “a Y

significant effect is produced by concentrations as low as 1 PM

with a maximum at 100 nM, allowing us to envisage that such an effect might be physiologically relevant if it were corrobo- rated in adult rats. In fact, during proestrus in female rats, plasma progesterone levels may rise to micromolar concentra- tions (Smith et al., 1975). Furthermore, allopregnanolone and allotetrahydrodeoxycorticosterone plasma levels rise during stress from concentrations lower than nanomolars to reach concentra- tions near of 10 nM (Purdy et al., 1991). Moreover, enzymes in the brain also can generate active metabolites from the parent ste- roids (i.e., progesterone), and local concentrations of the neuro- active steroids may reach high nanomolar concentrations because of the contribution from the local production. Thus, the proges- terone concentration in the brain of normal rats can oscillate between 5 and 22 nM (Purdy et al., 1991; Korneyev et al., 1993) and can increase up to 55-65 nM after electroshock (Korneyev et al., 1993) and up to 20-42 nM after swimming stress (Purdy et al., 1991). Allopregnanolone concentration in normal rat brain is 2-4 nM and can increase up to lo-20 nM after swimming stress (Purdy et al., 1991) or up to 4-9 nM after acute electroshock (Korneyev et al., 1993). Because of the range of concentrations of proges- terone and metabolites measured in plasma and brain, it seems reasonable to envisage that if the effect reported here is corrob- orated in vlvo in adult rats, it could have a physiological relevance in some of these particular conditions.

It is well established that local target tissue metabolism is an important factor in the mechanism of action of steroid hormones. 5oc-Dihydroprogesterone and allopregnanolone are the primary me- tabolites of progesterone accumulating in adult (Karavolas and Hodges, 1990) and fetal brain (Bamea et al., 1990). The site of formation of allopregnanolone is predominantly the brain (Corpe- chat et al., 1993), and the enzyme activities that catalyze these

picrotoxin (IO pM) +

500 3a, 5a-THP (10 nM) picrotoxin (IO PM) 3a, Sa-THP (IO nM)

1 I I 1 I I ,

250- "\-

3a, 5a-THP (10 nM) bicucullwz (50 I.IM) +

bicuculllne (50 PM) 3a, 5~THP (IO nM) I I I I

Figure 6. Effect of picrotoxin (A) and bicuculline (B) on 3cy,5a-THP-induced [Ca”], rise. A, B: Left, Control [Ca2+li responses induced by 10 nM allopregnanolone. Middle, When applied alone, 10 FM picrotoxin (A) and 50 WM bicuculline (B) did not significantly modify resting [CaZtli level. Right, [Ca2+li responses evoked by simultaneous application of 10 FM picrotoxin plus 10 nM allopregnanolone (A) and 50 PM bicuculline plus 10 nM allopregnanolone (B).

Dayanithi and Tapia-Arancibia l Allopregnanolone and [Ca2-‘li Rise in Hypothalamic Neurons J. Neurosci., January 1, 1996, 16(1):130-136 135

conversions are expressed strongly in the hypothalamus and anterior pituitary (Karavolas and Hodges, 1990). Both Sa-reductase and 3a-reductase activities are affected by the estrus cycle with peak activities at proestrus, estrus, and metes&us (Karavolas and Hodges, 1990). In addition, the hypothalamus and the anterior pituitary specifically accumulate significant amounts of radiolabeled Sa-dihy- droprogesterone and progesterone, the two substances being the precursors of allopregnanolone. [3H]progesterone also is metabo- lized to Sa-dihydroprogesterone and allopregnanolone in male rat hypothalamic slices (Korneyev et al., 1993). These data clearly show that the adult rat hypothalamus is equipped with the enzymes and substrates for the synthesis of neurosteroids, and these also may be modulated in viva by the ovarian cycle in the female.

Neurotransmitter and neurohormone release depend on cal- cium, and the modulation of calcium availability within the cell (either an increase in immature tissue or a decrease in adult tissue) may represent an important mechanism of action of active steroids at the cell surface. However, few reports are available on the regulatory effect of neurosteroids on calcium conductances or calcium concentrations in the central nervous system. Whole-cell patch-clamp studies in freshly isolated adult guinea pig hippocam- pal CA1 pyramidal neurons show that micromolar concentrations of allotetrahydrodeoxycorticosterone, dehydroepiandrosterone sulfate, and pregnanolone depress voltage-gated calcium currents (ffrench-Mullen and Spence, 1991). This effect also is reproduced by pregnenolone and pregnenolone sulfate but not by progester- one (Spence et al., 1991). This inhibition of calcium currents by the steroids seem to be mediated by a PTX-sensitive G-protein associated with the activation of PKC (ffrench-Mullen et al., 1994). But in cultured rat hippocampal neurons, high concentra- tions of pregnenolone sulfate failed to alter basal [Caztli or to modify the K+-induced [Ca’+], rise, as measured by microspec- trofluorimetry and the calcium-sensitive indicator Fura-2. How- ever, at these concentrations (5-250 FM), the steroid potentiated the NMDA-induced rise of [Cazfli in a dose-dependent manner (Irwin et al., 1992).

The possible functional significance that in developing neu- rons allopregnanolone may potentiate GABA-induced depo- larization leading to an increased Ca2+ influx is not clear. It is possible that it is involved in the development or modification of inhibitory synapses in hypothalamic circuits.

In conclusion, the present data are consistent with the presence of a receptor or recognition site for allopregnanolone at or close to the cell surface, probably associated to GABAA receptors, leading to changes in [Ca”‘], in hypothalamic neurons. These findings may open up entirely new areas of investigation, because rapid calcium changes induced by neurosteroids may represent a link of the metabolic cascade whereby hypothalamic neurons control the short-term process of release.

REFERENCES Barnea A, Hajibeigi A, Trant JM, Mason I (1990) Expression of

steroid metabolizing enzymes by aggregating fetal brain cells in culture: a model for developmental regulation of the progesterone Sa-reductase. Endocrinology 127:500-502.

. -

Baulieu EE (1981) Steroid hormones in the brain: several mechanisms? In: Steroid ho&one regulation of the brain (Fuxe K, Gustafsson JA, Wetterberg L, eds), pp 3-14. Oxford: Pergamon.

Baulieu EE (1991) Neurosteroids: a new function in the brain. Biol Cell 71:3-10.

Blackmore PF, Beebe SJ, Danforth DR, Alexander N (1990) Proges- terone and 17a-hydroxyprogesterone: novel stimulators of calcium in- flux in human sperm. J Biol Chem 265:1376-1380.

Cherubini E, Gaiarsa JL, Ben-Ari Y (1991) GAEiA: an excitatory trans- mitter in early postnatal life. Trends Neurosci 14:515-519.

Connor JA, Tseng HY, Hockberger PE (1987) Depolarization- and transmitter-induced changes in intracellular Ca*+ of rat cerebellar gran- ule cells in explant cultures. J Neurosci 7:1384-1400.

Corp&chot C, Young J, Calve1 M, Wehrey C, Veltz JN, Touyer G, Mouren M, Prasad VVK, Banner C, Sjiivall J, Baulieu EE, Robe1 P (1993) Neurosteroids: 3a-hydroxy-5a-pregnan-20-one and its precursors in the brain, plasma, and steroidogenic glands of male and female rats. En- docrinology 133:1003-1009.

Dayanithi G, Rage F, Richard P, Tapia-Arancibia L (1995) Character- ization of spontaneous and N-methyl-o-aspartate-induced calcium rise in rat-cultured hypothalamic neurons. Neuroendocrinology 61:243-255.

Desarmknien MG, Devic E, Rage F, Dayanithi G, Tapia-Arancibia L, Richard P (1994) Synchronous development of spontaneous and evoked calcium-dependent properties in hypothalamic neurons. Dev Brain Res 79:85-92.

Drouva SV, Laplante E, Kordon C (1985) Progesterone-induced LHRH release in vitro is an estrogen-as well as Ca’+--and calmodulin- dependent secretory process. Neuroendocrinology 40:325-331.

ffrench-Mullen JMH, Danks P, Spence KT (1994) Neurosteroids modulate calcium currents in hippocampal CA1 neurons via a pertussis toxin-sensi- tive G-protein-coupledmech&ism. J Neurosci 141963-1977.

ffrench-Mullen JMH. Suence KT (1991) Neurosteroids block Ca*+ chan- nel current in freshl; isolated hippoc\mpal CA1 neurons. Eur J Phar- macol202:269-272.

Grynkiewicz G, Poenie M, Tsien RY (1985) A new generation of Ca*+ indicators with greatly improved fluorescence properties. J Biol Chem 260:3440-3450.

Hofmann F, Biel M, Flockerzi V (1994) Molecular basis for Ca*+ chan- nel diversity. Annu Rev Neurosci 17:399-418.

Irwin RP, Lin SZ, Rogawski MA, Purdy RH, Paul SM (1994) Steroid potentiation and inhibition of N-methyl-D-aspartate receptor-mediated intracellular Ca*+ responses: structure-activity study. J Pharmacol Exp Ther 271:677-682.

Irwin RP, Maragakis NJ, Rogawski MA, Purdy RH, Farb DH, Paul SM (1992) Pregnelonone sulfate augments NMDA receptor mediated in- creases in intracellular Ca*+ m cultured rat hippocampal neurons. Neurosci Lett 141:30-34.

Karavolas HJ, Hodges DR (1990) Neuroendocrine metabolism of pro- gesterone and related progestins. In: Steroids and neuronal activity (Ciba Foundation Symposium 153, ed), pp 22-55. Chichester: Wiley.

Korneyev A, Guidotti A, Costa E (1993) Regional and interspecies dif- ferences in brain progesterone metabolism. JNeurochem 61:2041-2047.

Lambert RC. Davanithi G. Moos FC. Richard P (1994) Isolated suoraou- tic cells reipoid to oxyiocin by a’rise in intraiellular Ca2+ I ’ concentra- tion. J Physiol (Lond) 478:275-287.

Majewska MD, Harrison NL, Schwartz RD, Barker JL, Paul SM (1986) Steroid hormone metabolites are barbiturate-like modulators of the GABA receptor. Science 232:1004-1007.

McEwen BS (1991a) Steroid hormones are multifunctional messengers to the brain. Trends Endocrinol Metab 2:62-67.

McEwen BS (1991b) Non-genomic and genomic effects of steroids on neural activity. Trends Pharmacol Sci 12:141-147.

Mellon SH (1994) Neurosteroids: biochemistry, modes of action, and clinical relevance. J Clin Endocrinol Metab 78:1003-1008.

Orchinik M, McEwen B (1993) Novel and classical actions of neuroactive steroids. Neurotransmissions 9:1-6.

Park-Chung M, Wu FS, Farb DH (1994) 3a-Hydroxy-SP-pregnan- 20-one sulfate: a negative modulator of the NMDA-induced current in cultured neurons. Mol Pharmacol 46:146-150.

Paul SM, Purdy RH (1992) Neuroactive steroids. FASEB J 6:2311-2321. Purdy RH, Morrow AL, Moore PH, Paul SM (1991) Stress-induced

elevations of y-aminobutyric acid type A receptor-active steroids in the rat brain. Proc Nat1 Acad Sci USA 88:4553-4557.

Rage F, Benyassi A, Arancibia S, Tapia-Arancibia L (1992) y-Ami- nobutyric acid-glutamate interaction in the control of somatostatin release from hypothalamic neurons in primary culture: in vivo cor- roboration. Endocrinology 130:1056-1062.

Rage F, Jalaguier S, Rougeot C, Tapia-Arancibia L (1993) GABA inhibition of somatostatin gene expression in cultured hypothalamic neurons. NeuroReport 4:320-322.

Ramirez VD, Dluzen DE, Ke FC (1990) Effects of progesterone and its metabolites on neuronal membranes. In: Steroids and neuronal

136 J. Neurosci., January 1, 1996, 16(1):130-136 Dayanithi and Tapia-Arancibia . Allopregnanolone and [Ca’ ‘I, Rise in Hypothalamic Neurons

activity (Chadwick D, Widdows K, eds), pp 125-144. Chichester: Wiley.

Robe1 P, Beaulieu EE (1990) Les neurosteroides: une nouvelle fonc- tion du cerveau? Medicine Sciences 6:252-260.

Robe1 P, Beaulieu EE (1994) Neurosteroids: biosynthesis and func- tion. Trends Endocrinol Metab 5:1-8.

Schumacher M (1990) Rapid membrane effects of steroid hormones: an emerging concept in neuroendocrinology. Trends Neurosci 13:359-362.

Simpson PB, Challiss RAJ, Nahorski SR (1995) Neuronal Ca2+ stores: activation and function. Trends Neurosci 18:299-306.

Smith MS, Freeman ME, Neil1 JD (1975) The control of progesterone secretion during the estrous cycle and early pseudopregnancy in the rat: prolactin, gonadotropin and steroid levels associated with rescue on the corpus Iuteum of pseudopregnancy. Endocrinology 96:219-226.

Smith SS, Waterhouse BD, Chapin JK, Woodward DJ (1987) Proges- terone alters GABA and glutamate responsiveness: a possible mech- anism for its anxiolytic action. Brain Res 400:353-359.

Spence KT, Plata-Salaman CR, ffrench-Mullen JMH (1991) The neurosteroids pregnenolone and pregnenolone-sulfate but not pro- gesterone, block Ca2+ currents in acutely isolated hippocampal CA1 neurons. Life Sci 49:235-239.

Tapia-Arancibia L, Humbert T (1991) Activation of dihydropyridine- sensitive calcium channels induces somatostatin release from hypo- thalamic neurons. Pharmacological characterization. Neurochem Int l&367-371.

Tapia-Arancibia L, Pares-Herbute N, Astier H, Reichlin S, Nathanson J (1988) Adenylate cyclase activation is not sufficient to stimulate somatostatin release from dispersed cerebral cortical and dience- phalic cells in glia-free cultures. Brain Res 450:101-110.

Wu FS, Gibbs TT, Farb DH (1991) Pregnenolone sulfate: a positive allosteric modulator at the N-methyl-o-aspartate receptor. Mol Pharmacol 40:333-336.

Yuste R, Katz LC (1991) Control of postsynaptic Ca2+ influx in de- veloping neocortex by excitatory and inhibitory neurotransmitters. Neuron 6:333-344.